Deleting users from the OJS database

Software Support
42

hi @obi,

it is an alias for table article , the same is for r and table roles…

regards

43

I can’t see any table in the database called articles in OJS 3.x

44

@obi, sorry, I was playing on an OJS2.x database… in OJS 3 not even roles table exist!

on OJS3 you can use tables users and authors to figure out user who are not in authors table…
also user_groups to see if they have other roles (i guess)

45

I made a username list (in ojs2) by exporting list of URL column and making frequency list (some urls was in 20-30k user profiles) and by exporting usernames wits “stop phrases” in user_settings table.

Now, I deleted about 300K users (via mergeUsers.php, it is enough slow) from this list (only with reader role) and I see about 200K users in my system now. It is no more possible to delete bad users by big groups (there is no stop phrases or bad urls more).

ReCAPTCHA is enough good. It was about 1000 new users per day, and now I see about 10 new users per day.

One more strange moment - more than 90% of bad users was enrolled in one journal (we have 15 journals in our ojs). There was no comments or articles by this authors, but profiles (with porn, casino and other) was in google cache.

Conspiracy theory advocates say this is an “attack” on the journal.

46

sorry to be so basic, but how do I bring up a command line if on a home pc for this equivalent: https://JournalURL/base/tools/mergeUsers.php

47

Hi @gail,

That depends on your operating system. Have a look e.g. at websocket - How to run php file through command line? - Stack Overflow.

Regards,
Alec Smecher
Public Knowledge Project Team

48
  1. I have a php executable in my “path” on windows and on linux.
  2. I make a list of “bad” users by database query ( for example select username from ojs.users WHERE user_id in (select user_id from ojs.user_settings WHERE setting_value like '%porno%');)
  3. I make a batch file in a text editor by replacing “\nbadusername” to “\nphp pathtotools\megreusers.php tempusername badusername”
  4. After this, I just run batch file in the shell (windows or linux) and see a result. Deleting/merging of big (10K+) groups of users is very long procedure.
3 Likes
[OJS 3.1.2.1] Cannot add any existing user to a new journal
50

Hi.

Help me please to make an sql querry, to get a list of users with their registration date, last login date and group. May be, it will be good even to get only users (usernames) with only a “reader” role with registration date = last login date.

I want to delete all “readers only” with registration and last login before 2019.

I allready deleted more than 750K bots, bul last 20-30K are not clearable in batch mode.

51

Hi. We made a great cleanup of our database. A lot of users was deleted/megred. But it is unclear for me how to cleanup “user_interests” table in right way. We have a lot of allready deleted users (ids) in this table.

May be, mergeUsers.php should be impoved?

1 Like
52

Can you advise me if I’m forming the SQL query correctly to delete users from my database?

DELETE FROM users
WHERE id IN (
  SELECT user_id
  FROM comments
  WHERE content LIKE '%casino%'
);

Also, does anyone know how to prevent new spam users from simply registering? I understand how this can be done when they use direct words like “casino” in their spam, but not every link that is spam contains these words.

1 Like
Bulk Deletion of Fake "Reader" Users in OJS 3.2.1.2 with PostgreSQL
55

I’m a journal editor using the OJS through another institution managing the platform on our behalf. It is not clear to me how deleting a user works using the merge users feature. My concern is for EU users who wish to be forgotten/erased under the GDPR. Is there a way for users to delete themselves using the merge feature? If so, how, please, since when you click on it, it appears to require that you choose another record with which to merge the one you’re in? I need to be able to post directions for our EU readers, even if the directions are simply that they must contact the editors to be forgotten (which seems annoying both for the user and the editor–it would be much preferable A) for users to be able to control their own records, and B) for the language around that control to be far more clear than “merge”). Is the best answer for me to ask our host institution to handle such cases?